A novel computational cognitive model explains human procedural error in terms of declarative memory processes. This is an early version of a process model intended to predict and explain multiple classes of procedural error a priori. We begin with postcompletion error (PCE), a type of systematic procedural error that people are prone to commit when there is one step to perform after they have accomplished their main task goal. Participants in an experiment demonstrated increased PCE rates following an interruption in a realistic form-filling task. The model explains PCE as a consequence of two declarative retrieval processes, spreading activation and base-level activation, competing with each other because of features of task and working memory structure. Our intention is to generalize the model to other classes of procedural error in complex task environments.